Detection And Tracking Of Infrared Criminal Investigation Target Based On Artificial Immunity Algorithm

With the development of science and technology,target detection and tracking technology has been gradually applied in criminal investigation.Compared with the traditional detection method of criminal investigation,the method combined with image processing technology broadens the direction of solving cases.In the investigation of the case,the method of collecting the scene traces by infrared equipment can locate the traces quickly according to the residual temperature information on the spot.And on the other hand,the traces which can not be detected under visible light can be observed in infrared image,which is conducive to providing effective supporting evidence for solving case.However,in the actual acquisition process of infrared criminal investigation image,the complex background and target structure easily make the shape characteristics of infrared target different from the actual characteristics,and the interference of environmental factors such as occlusion of the target easily makes the target characteristics incomplete.In addition,due to the heat transfer between the infrared target and the environment,the edge of the target is prone to blur.These problems will affect the reliability and accuracy of target extraction and detection.Therefore,the design of effective target extraction and detection methods for infrared criminal investigation images with complex background and blurred edges,and then the accurate tracking of the target,plays an important role in improving the detection efficiency of criminal cases.This thesis focuses on the detection and tracking of different criminal investigation targets,including:(1)Aiming at the uneven distribution of target features in infrared criminal investigation images,an infrared target detection algorithm combining clonal threshold and growing immune field is proposed.The proposed algorithm uses the mechanism of biological immunity to design the growing immune field,and solves the problem of over-growth of traditional region growing algorithm.(2)In order to improve the accuracy for complex infrared criminal investigation image detection algorithm,an infrared image target extraction algorithm with adaptive growth immune field is proposed.The problems of edge blurring,similar target interference,complex background features and significant difference of targets in infrared criminal investigation images under complex background are analyzed.Aiming at these problems,the design and algorithm improvement of adaptive growth immune field is realized based on the region information and edge gradient information of image.(3)Aiming at the problem of infrared handprint extraction with discrete feature distribution and blurred edge,an edge blurred infrared target extraction algorithm based on geometric growth with structural morphology is proposed.The algorithm uses the maximum entropy method and threshold expansion to segment the image,extracts the feature points of the region block from the target region to construct the structure of the hand.Seed points are searched by region block feature points,and the distance relationship between seed points and corresponding feature points is used as the judgment condition of geometric growth,and finally the target image is obtained.(4)Aiming at the occlusion problem of criminal investigation target in different moving states,uniform velocity criminal investigation target tracking algorithm based on motion estimation and non-uniform velocity criminal investigation target tracking algorithm combined with background difference and motion estimation are proposed respectively.By analyzing the motion vectors of the target,the moving position of the target is predicted,and the accuracy of target location detection is improved by using the background difference method.In summary,referring to the biological immune mechanism,this thesis designs the infrared criminal investigation target detection and tracking algorithms based on artificial immunity.Through comparative experiments,they show excellent detection and tracking effect according to the characteristics of each criminal investigation target.